[PDF] Comparisons in Performance of Electromagnet and Permanent

29159_320100035731.pdf

Comparisons in Performance of Electromagnet and

Permanent-Magnet Cylindrical Hall-Effect Thrusters

Kurt A. Polzin*

NASA-Marshall Space Flight Center, Huntsville, AL 35812 Yevgeny Raitses†, Jean Carlos Gayoso and Nathaniel J. Fisch$ Princeton Plasma Physics Laboratory, Princeton, NJ 08543

Three different low-power cylindrical Hall thrusters, which more readily lend themselves to miniatur-

ization and low-power operation than a conventional (annular) Hall thruster, are compared to evaluate the

propulsive performance of each. One thruster uses electromagnet coils to produce the magnetic field within

the discharge channel while the others use permanent magnets, promising power reduction relative to the elec-

tromagnet thruster. A magnetic screen is added to the permanent magnet thruster to improve performance by

keeping the magnetic field from expanding into space beyond the exit of the thruster. The combined dataset

spans a power range from 50-350 W. The thrust levels over this range were 1.3-7.3 mN, with thruster effi-

ciencies and specific impulses spanning 3.5-28.7% and 400-1940 s, respectively. The efficiency is generally

higher for the permanent magnet thruster with the magnetic screen, while That thruster's specific impulse as

a function of discharge voltage is comparable to the electromagnet thruster.

I. Introduction

W

HILE annular Hall thrusters can operate at high efficiency at kW power levels, it is difficult to construct one that

operates over a broad envelope from ∼ 1 kW down to ∼ 100 W while maintaining an efficiency of 45-55%. Scal-

ing to low power while holding the main dimensionless parameters constant requires a decrease in the thruster channel

size and an increase in the magnetic field strength. 1,2 Increasing the magnetic field becomes technically challenging

since the field can saturate the miniaturized inner components of the magnetic circuit and scaling down the magnetic

circuit leaves very little room for magnetic pole pieces and heat shields. In addition, the central magnetic pole piece

defining the interior wall of the annular channel can experience excessive heat loads in a miniaturized Hall thruster,

with the temperature eventually exceeding the Curie temperature of the material 2 and in extreme circumstances leading

to accelerated erosion of the channel wall.

An alternative approach is to employ a cylindrical Hall thruster (CHT) geometry. 3 Laboratory model CHTs have

operated at power levels ranging from - 50 W up to - 1 kW. These thrusters exhibit performance characteristics that

are comparable to conventional, annular Hall thrusters of similar size. Compared to the annular Hall thruster, the

CHT's insulator surface area to discharge chamber volume ratio is lower. Consequently, there is the potential for

reduced wall losses in the channel of a CHT, and any reduction in wall losses should translate into lower channel

heating rates and reduced erosion, making the CHT geometry promising for low-power applications. This potential

for high performance in the low-power regime has served as the impetus for research and development efforts aimed

at understanding and improving CHT performance. 3-7

In this paper, we present performance measurements from three separate variations on the CHT. Each variation is

different in the manner in which the magnetic field is applied to the discharge channel. The use of electromagnets in

one thruster offers the capability to test over a wide range of magnetic field strengths. The thrusters fabricated with

permanent magnets do not offer this flexibility, but they do have the promise of reduced overall power consumption

relative to the electromagnet variant. The use of permanent magnets simplifies the design by removing the multi-turn

*Propulsion Research Engineer, Propulsion Research and Technology Applications Branch, Propulsion Systems Department. Senior Member

AIAA. tResearch Physicist. Associate Fellow AIAA. *Professor, Astrophysical Sciences Dept. Senior Member AIAA.

1 of 7

American Institute of Aeronautics and Astronautics

A) Ceramic

Electromagnets Cha)nnel

B

Anode Annular Cathode

Region Neutralizer

electromagnet coils from the thruster. Thruster performance measurements (thrust, I sp, and anode efficiency) and the

discharge current as a function of voltage are presented to compare the performance of the different thrusters.

II. Experimental Apparatus

All cylindrical Hall thruster testing presented in this paper was conducted at NASA's Marshall Space Flight Center

(MSFC). We proceed first with a description of the three different thrusters and then discuss the facilities at MSFC.

A. CHT with Electromagnet Coils

Performance measurements obtained for the 3-cm electromagnet CHT (CHTem) shown in Fig. 1 were previously

reported in Ref. [6]. The thruster consists of a boron-nitride ceramic channel, an annular anode, two electromagnet

coils, and a magnetic core. The thruster channel is a composite of a shorter, annular region and a longer, cylindrical

region. Gas is injected through the anode into the short, annular region of the thruster. The length of the annular region

is sized to be greater than the ionization mean free path for xenon. This allows for high ionization of the propellant at

the boundary between the annular and cylindrical regions. The electromagnet coils are operated using independently

controlled power supplies, and the resulting field topology has a mirror-like structure near the thruster axis.

Figure 1. A) Schematic illustration and B) photograph of the 3-cm electromagnet cylindrical Hall thruster.

B.CHT with Permanent Magnets

Various measurements on the 2.6-cm channel diameter permanent magnet CHT (CHTpm-1) shown in Fig. 2a were

previously reported in Refs. [8, 9]. The thruster is roughly 5.5 cm in overall diameter and 3.5 cm long, massing

roughly 350 g. The thruster channel is comprised of a ceramic boron-nitride insulator with propellant fed from an

annular anode. Unlike the CHTem, this thruster is purely cylindrical, possessing no shorter, annular region. Two sets

of samarium-cobalt (Sm-Co) rare-Earth magnets are used to produce the magnetic field. The magnets are oriented in

the same direction to produce the 'direct' magnetic field topology shown in Fig. 2b and the maximum field strength

inside the thruster channel is roughly 1 kG. The cathode is oriented such that the orifice is 1.9 cm (0.75") downstream

of the thruster exit and 5.1 cm (2") from the thruster centerline. The cathode is aligned such that its centerline forms a

40o angle with the thruster centerline.

C.CHT with Permanent Magnets and a Magnetic Screen

Recent measurements were obtained on the 2.6-cm channel diameter permanent magnet CHT (CHTpm-2) shown in

Fig. 3. This is the same thruster that described in the previous section, except it additionally possesses an external

magnetic screen to keep the magnetic field from expanding into space downstream of the thruster. The thruster is

operated with the cathode in one of two positions, with the orifice radially offset 1.47 cm in the minimum radial

position and 2.35 cm in the 'zero' position. In both cases, the cathode is oriented perpendicular to the thruster axis and

the orifice is located roughly 2 cm downstream of the thruster exit plane. 2of7 American Institute of Aeronautics and Astronautics a) b) a) 1.47 b) 2.35

Figure 2. a) Laboratory model 2.6-cm CHT with Sm-Co permanent magnets (with US quarter for scale). b) Magnetic field topology in a

laboratory model 2.6 cm CHT with Sm-Co permanent magnets where the magnets are oriented in a direct-field alignment. The maximum

magnetic field is roughly 1 kG at the axis near the back wall.

Figure 3. Laboratory model 2.6-cm CHT with Sm-Co permanent magnets and a magnetic screen outer shell with the cathode a) in the

minimum radial position and b) in the 'zero' position. c) Magnetic field map of the thruster.

D. NASA-MSFC Test Facility

In all three cases a commercial HeatWave Labs HWPES-250 hollow cathode is used, serving as both the thruster

cathode and the beam neutralizer. The working propellant for all experiments is research-grade xenon gas, and the

cathode and anode flow rates are independently controlled. The propellant flow rate to both the cathode and anode

were controlled using two variable 10-sccm MKS 1479 precision flow controllers (calibrated on Xe and controllable

to f0.1 sccm). All testing was performed with a cathode flow rate of 2 sccm. Power was provided to the thruster and

cathode using Sorenson switching power supplies.

Testing was conducted in a 2.75-m diameter, 7.6-m long stainless steel vacuum chamber. The vacuum level inside

the chamber is maintained by two 9500l/s gaseous helium cryopumps. The base pressure of the facility was 2 × 10 - 7

torr, and the pressure level (corrected for Xe) during testing was roughly 1-2× 10 - 5 torr.

Thrust was measured using the variable-amplitude hanging pendulum with extended range (VAHPER) thrust

stand.10 The stand employs a unique linkage mechanism to convert horizontal deflection of the pendulum arm into

amplified vertical deflection of a secondary beam. Displacement (thrust) calibration of the VAHPER thrust stand is

accomplished using an in situ calibration rig that applies a series of known loads normal to the pendulum arm. Cal-

3 of 7

American Institute of Aeronautics and Astronautics a) . le^A n_ •

CHTpm-2

anode f'ow (sccm) min radius 'zero' • 3.0 o 3.0 n 3.5 q 3.5. 4.0 0 4.0 d i so^ge,c .p o w er (W o .a o•NAA o A& o ^q • d) e) c a^ a^ 8 7 6 5 4 3 2 1 0 30
25
20 15 10 5 0 c) 2000 1500
1000
500
0 f) 2000 1500
1000
500
0 d i s c h a r g e p o ^• w e r ( Wo iVOO•

0W°

ibration can be performed before, during, and after thruster operation. The measured displacement of the vertically

deflecting linkage is recorded as the calibration loads are applied to the arm. Assuming that the relationship between

the applied force and the measured displacement is linear allows for a linear curve fit of the calibration data.

It should be noted that there are still some issues that remain to be explored, especially with CHTpm-2. The most

important is the fact that, for a given discharge current, the thruster draws more current in the MSFC setup than in the

test setup at the Princeton Plasma Physics Laboratory (PPPL). It is unclear at this time how important this effect is

with respect to the integrity of the measured data. It also needs to be explored in terms of whether the source of this

discrepancy resides in the test setup (cathode mount position, etc.) or is a facility effect.

III. Experimental Data

Thrust, thruster efficiency, and specific impulse (Isp) measurements for all the thrusters described in the previous

section are presented in Fig. 4. The thrust and specific impulse measurements are given according to their standard

definitions.11 To account for the added power required by the CHTem relative to the CHTpm models, the thruster

efficiency is calculated as the ratio of the exhaust jet power to the sum of the discharge power and the electromagnet

power. For the CHTem, the power applied to the electromagnet coils is roughly 100 W. CHTpm-1 and -2 are not

penalized by this additional power loss because they do not have electromagnet coils. The data for CHTem and

CHTpm-1 are displayed in Figs. 4a-c, while the data for CHTpm-2 operating with the cathode exit located at the

minimum radius and 'zero' positions are shown in Figs. 4d-f. b) c a^ a^ 8 7 6 5 4 3 2 1 0

° o® oIsv^

A

FPO anode f'ow (sccm)

o CHTem CHTpm-1 • 3.0 o 3.4 n 3.9 q 4.0® 0 4.4 v 5.0

30 0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 0

25 d i s c h a r g e p o w e r ( W )

20 .,^, ^p ° 8 0

15 o

10 ro®a o5

0

0 50 100 150 200 250 300 350 0 50 100 150 200 250 300 350

discharge power (W) discharge power (W)

Figure 4. Thrust, thruster efficiency (including electromagnet power consumption for CHTem), and IsP as a function of discharge power

for a)-c) the CHTem and CHTpm-1 and d)-f) the CHTpm-2 for both cathode positions (minimum radius and 'zero'). The average error

bars (with a 95% confidence interval) on these data are roughly 150 µN,1.1%, and 40-50 sec.

4 of 7

American Institute of Aeronautics and Astronautics

100 200 300 400 500

discharge voltage (V)

0.8CHTpm-1 anode flow (sccm)

0.7- o 3.4 o 4.00 4.4 v 5.0V

0.6 ° o

V 0

0.5 q o ^A

® q0 00.4 ®0@00 0

0.3 b) Q ca^ a^rn m 'v

0 nn 3.9 sccm

g I @ $ • 3.0 sccm d) Q ca^ a^rn m 'v 0.8 0.7 0.6 0.5 0.4 0.3 CHTem a) Q c a^rn m w'v 0.8 0.7 0.6 0.5 0.4 0.3 c) Q c a^rn m w'v 0.8 0.7 0.6 0.5 0.4 0.3

The data span the range from 50-350W in discharge power, resulting in thrust levels between 1.3 and 7.3 mN,

thruster efficiencies between 3.5 and 28.7%, and Isp levels between 400 and 1940 sec. Thrust and specific impulse

increase with discharge power in all cases. Thruster efficiency increases with discharge power below 200 W, but in the

instances where data are available the efficiency starts to asymptote or even decrease above 200 W.

The performance data show that the CHTpm-2 with the magnetic screen installed generally has greater efficiency

for both cathode positions. Below 150 W with the cathode at the minimum radius position the efficiency at a given

discharge power level is roughly the same for all three flow rates tested. Just as interesting is that the efficiency at

the low power level of 50 W and a flow rate of 3.0 sccm is 9.9% While the CHTem has lower efficiency relative to

CHTpm-2, it should be noted that the thrust and Isp of the former are higher and it is only the electromagnet power that

reduces the efficiency of the CHTem to the level reported here. At low discharge power levels, there is a higher degree

of variability and spread in the specific impulse and efficiency of CHTpm-1 relative to the other data sets. During

testing the operation of CHTpm-1 at lower discharge power levels was much more sensitive to very small changes in

the discharge parameters relative to the other thrusters, producing the inconsistent or scattered results observed in the

data set. ♦ ♦ CHTpm-2 ♦ ♦ min radius n ♦ ♦ 4.0 sccm♦ n n n n 3.5 sccm • • • _ n n; • 3.0 sccm

100 200 300 400 500

discharge voltage (V)

0 o CHTpm-2

0A '%ero'

o q A 4.0 sccm

0 00 0 q 3.5 sccmq o

0 0 3.0 sccm

100 200 300 400 500 100 200 300 400 500

discharge voltage (V) discharge voltage (V)

Figure 5. Discharge current as a function of discharge voltage and anode flow rate for a) CHTem, b) CHTpm-1, c) CHTpm-2 (minimum

radius cathode position), and d) CHTpm-2 ('zero' cathode position).

The discharge current as a function of voltage for all thruster configurations is presented in Fig. 5. For a given

thruster, we observe that in all cases the current is greater at higher flow rates. While only a small span of discharge

voltages was tested using the CHTem, the data over this span exhibit a relatively constant discharge current. Except

for the lowest flow rate, the discharge current for CHTpm-1 exhibits a local minimum between 250 and 300 V. While

the current levels for CHTem and CHTpm-1 are somewhat comparable, the currents drawn by CHTpm-2 represent

a significant increase. Functionally, both CHTpm-2 data sets follow the same basic form, exhibiting a local maxima

at lower voltages. While the discharge in the CHTpm-2 extinguished above 350 V in the 'zero' configuration, the

other cathode position did permit operation at higher voltages and resulted in the appearance of local minima in the

discharge current.

The Isp as a function of discharge voltage is given in Fig. 6 for all thruster configurations. In all cases, I sp increases

with discharge voltage. In most cases, Isp is higher at higher flow rates. However, for CHTpm-2 with the cathode

in the 'zero' position, the lowest specific impulse at a given discharge voltage occurs at 3.5 sccm, with higher I sp at

both 3.0 and 4.0 sccm. From 250-350 V, the specific impulse of CHTem and CHTpm-2 are comparable in magnitude.

5 of 7

American Institute of Aeronautics and Astronautics b) 2000 1500
1000
500
a) 2000 1500
1000
500
0 0 see `^^qanode flow (sccm)

8 0 ®0 CHTem CHTpm-1

0 • 3.0 0 3.4

0 n 3.9 q 4.00 4.4v 5.0

• • CHTpm-2 •q anode flow (sccm) q min radius 'zero' • 3.0 o 3.0n 3.5 q 3.5• 4.0 0 4.0

100 200 300 400 500 100 200 300 400 500

discharge voltage (V) discharge voltage (V)

Figure 6. IeP as a function of discharge voltage for a) the CHTem and CHTpm-1, b) the CHTpm-2 for both cathode positions (minimum

radius and 'zero').

The effect of the magnetic screen on the permanent magnet thruster performance is clearly seen in the upwards shift

in specific impulse for CHTpm-2 relative to CHTpm-1 at a given discharge voltage.

For those cases where at a given voltage Isp is greater for a given mass flow rate, we can speculate that this increase

is due to greater propellant use (i.e., low neutral fraction), the formation of doubly ionized Xe, or a combination of the

two. However, we do not have enough information at this time to reach a definite conclusion.

IV. Conclusions

Although conventional (annular) Hall thrusters are efficient in the kilowatt power regime, they become inefficient

when scaled to small sizes. This is due to the difficulties associated with holding the performance scaling parameters

constant while decreasing the channel size and increasing the applied magnetic field strength. The cylindrical Hall

thruster can be more readily scaled to smaller sizes due to its nonconventional discharge-chamber geometry and

associated magnetic field profile.

In the present paper, a series of measurements obtained on three separate cylindrical Hall thrusters were compared.

One thruster employed an electromagnet coil to generate the magnetic field in the thruster while the other two used

permanent magnets. The use of permanent magnets promises lower thruster power consumption by eliminating the

electromagnet power sink. In addition, permanent magnets simplify the design, removing the need for additional

power supplies and multi-turn electromagnet coils.

Performance measurements for all thrusters were obtained using the VAHPER thrust stand. The combined data

span the range of discharge power from 50-350 W. The thruster produced thrust levels ranging from 1.3-7.3 mN,

thruster efficiencies spanning 3.5-28.7%, and Isp between 400-1940 sec. All performance parameters generally in-

creased with discharge power, except in testing of the CHTpm-1 above 250 W. For that thruster above 250 W, the

thruster efficiency was reduced with increasing discharge power. The performance data show a generally higher ef-

ficiency for CHTpm-2 for both cathode positions, with the capability to operate at lower power levels. The I sp as a

function of discharge voltage was comparable between CHTem and CHTpm-2, while the specific impulse on CHTpm-

1 was lower, indicating the positive effect of the magnetic screen on the performance of the permanent magnet thruster.

Acknowledgments

Work performed by the Princeton Plasma Physics Laboratory coauthors was partially supported by the U.S. Air

Force Office of Scientific Research. We gratefully acknowledge the contributions of Enrique Merino, who contributed

technical assistance in the design of the magnetic screen, and to MSFC technical support staff Tommy Reid and Doug

Galloway. We extend our thanks to students Adam Kimberlin, Ryan Sullenberger, Valerie Hanson, Lindsey Walker,

and Kevin Bonds for their efforts. We also appreciate and acknowledge the continued MSFC support of Mr. James

Martin, Mr. J. Boise Pearson, Dr. Thomas Brown, and Mr. Thomas Williams. 6of7 American Institute of Aeronautics and Astronautics

References

1A.I. Morozov and V.V. Savelyev, "Fundamentals of Stationary Plasma Thruster Theory," Reviews ofPlasma Physics, edited by F.F. Kadomtsev

and V.D. Shafranov, Vol. 21, Consultants Bureau, New York, 2000, p. 203.2V. Khayms and M. Martinez-Sanchez, "Fifty-Watt Hall Thruster for Microsatellites," Micropropulsion for Small Spacecraft, edited by M.M.

Micci and A.D. Ketsdever, Vol. 187, Progress in Astronautics and Aeronautics, AIAA, Reston, VA, 2000, pp. 233-254.3Y. Raitses and N.J. Fisch, "Parametric Investigations of a Nonconventional Hall Thruster," Phys. Plasmas, 8(5):2579, 2001.

4A. Smirnov, Y. Raitses, and N.J. Fisch, "Parametric Investigations of Miniaturized Cylindrical and Annular Hall Thrusters," J. Appl. Phys.,

92(10):5673, 2002.

5A. Smirnov, Y. Raitses, and N.J. Fisch, "Plasma Measurements in a 100 W Cylindrical Hall Thruster," J. Appl. Phys., 95(5):2283, 2004.

6K.A. Polzin, T.E. Markusic, B.J. Stanojev, A. Dehoyos, Y. Raitses, A. Smirnov, and N.J. Fisch, "Performance of a Low-Power Cylindrical

Hall Thruster," J. Propuls. Power, 23(4):886, 2007.

7Y. Raitses, E. Granstedt, A. Smirnov, E. Merino, and N.J. Fisch, "Effects of Cathode Electron Emission on Hall Thruster Discharge," AIAA

Paper 2008-5188, July 2008.

aK.A. Polzin, E.S. Sooby, A.C. Kimberlin, Y. Raitses, E. Merino, and N.J. Fisch, "Performance of a Permanant-Magnet Cylindrical Hall-Effect

Thruster," AIAA Paper 2009-4812, Aug. 2009.

9K.A. Polzin, E.S. Sooby, Y. Raitses, E. Merino, and N.J. Fisch, "Discharge Oscillations in a Permanent Magnet Cylindrical Hall-Effect

Thruster," 31st International Electric Propulsion Conference, Ann Arbor, MI, IEPC Paper 2009-122, Aug. 2009.

10K.A. Polzin, T.E. Markusic, B.J. Stanojev, A. DeHoyos and B. Spaun, "Thrust Stand for Electric Propulsion Performance Evaluation," Rev.

Sci. Instrum., 77:105108, 2006.11R.G. Jahn, Physics ofElectric Propulsion, McGraw-Hill, New York, 1968, pp. 4, 7.

7of7 American Institute of Aeronautics and Astronautics